Quick and easy access to the required information is important for maintenance staff to perform operations efficiently and to minimise system downtime. The integration of augmented reality (AR) into digital twin systems can improve access to such information. However, the development time, cost, and scalability challenges of AR applications have hindered their widespread adoption in industry. This paper presents a digital twin system architecture that integrates AR and knowledge graphs to dynamically update the user interface of the AR application as new information becomes available. The architecture combines a digital twin hierarchy, a services network, a knowledge graph, and AR applications. A preliminary evaluation indicates the architecture’s potential as a scalable, cost-effective, and adaptable solution for providing tailored, context-specific information to maintenance technician workflows.